Toggle light / dark theme

Now We’re Talking: NVIDIA Releases Open Dataset, Models for Multilingual Speech AI

The new Granary dataset was used to train high-accuracy and high-throughput speech AI models for audio transcription and translation.

/* */